One-year follow-up of the CAPSID randomized trial for high-dose convalescent plasma in severe COVID-19 patients

BACKGROUND Results of many randomized trials on COVID-19 convalescent plasma (CCP) have been reported, but information on long-term outcome after CCP treatment is limited. The objectives of this extended observation of the randomized CAPSID trial are to assess long-term outcome and disease burden in patients initially treated with or without CCP. METHODS Of 105 randomized patients, 50 participated in the extended observation. Quality of life (QoL) was assessed by questionnaires and a structured interview. CCP donors (n = 113) with asymptomatic to moderate COVID-19 were included as a reference group. RESULTS The median follow-up of patients was 396 days, and the estimated 1-year survival was 78.7% in the CCP group and 60.2% in the control (P = 0.08). The subgroup treated with a higher cumulative amount of neutralizing antibodies showed a better 1-year survival compared with the control group (91.5% versus 60.2%, P = 0.01). Medical events and QoL assessments showed a consistent trend for better results in the CCP group without reaching statistical significance. There was no difference in the increase in neutralizing antibodies after vaccination between the CCP and control groups. CONCLUSION The trial demonstrated a trend toward better outcome in the CCP group without reaching statistical significance. A predefined subgroup analysis showed a significantly better outcome (long-term survival, time to discharge from ICU, and time to hospital discharge) among those who received a higher amount of neutralizing antibodies compared with the control group. A substantial long-term disease burden remains after severe COVID-19. Trial registration EudraCT 2020-001310-38 and ClinicalTrials.gov NCT04433910. Funding Bundesministerium für Gesundheit (German Federal Ministry of Health).


INTRODUCTION
comparison based on the cumulative amount of transfused neutralizing units ( Fig. 2B and 2C). The better 153 outcome of the subgroup which has received a higher cumulative amount of neutralizing units was 154 confirmed in the final data set including long-term observation. It shows a significant shorter time to first 155 negative SARS-CoV-2-PCR (p=0.02), a shorter time to discharge from ICU (p=0.02) and discharge from 156 hospital (p=0.02) (log rank test, Fig. 3 A to D). The primary outcome of the study, i.e. survival and no longer 157 fulfilling criteria of severe COVID-19 on day 21, remained non-significant. In the final data set, among 158 those who received a high or low cumulative amount of neutralizing units, the primary outcome occurred 159 in 56.0% and in 32.1%, and in 30.8% in the control group (p=0.046 high titer vs. control). 160 161

Medical events during long-term follow-up 162
Patients reported GI symptoms (including abdominal pain, diarrhea, nausea, weight loss), pulmonary 163 symptoms, dyspnea pain symptoms, confusion, dizziness, hypersomnia, insomnia, conjunctivitis or 164 alopecia ( Table 2). The control group patients reported numerically less often gastrointestinal, or pain 165 symptoms than the CCP group (p=n.s.) Pulmonary symptoms were reported in 47% of patients in the CCP 166 group and 70% of patients in the control group (p=0.15) and during extended follow-up supplemental 167 oxygen was needed in 10% of patients in the CCP group but in 30% of patients in the control group 168 (p=0.13). During the extended follow-up period, 18% of patients were hospitalized and 18% of patients 169 needed supplemental oxygen. Twenty percent of patients in the CCP group and 15% of the control group 170 were hospitalized (p=0.724). The duration of hospitalization in the CCP group was 5 (3-6) days compared 171 to 15 (6-77) days in the control group (p=0.09). The proportion of hospitalization did not significantly differ 172 between patients who have received a high cumulative amount of neutralizing units compared to those 173 treated with a low cumulative amount of neutralizing units (6.2% vs. 35.7%, p=0.07). The use of radiologic 174 imaging of the chest was comparable between all groups. 175 Functional limitations assessed by the post COVID-Scale (i.e. grade 1 to 4) were reported by 56% of 176 patients (Fig. 4 A). Grade 2-4 functional limitations were reported by 48% of patients . The number of 177 patients reported to be free of limitations was not significantly different between CCP group (53%) and 178 control group (30%) (p=0.136) (Fig. 4 A). 179 Any medical event during follow-up was reported in 73% of donors and 84% of patients. Events rated as 180 grade 3 or higher occurred in 8% of donors and in 22% of patients (p=0.018). In donors, the most frequent 181 symptoms were neurologic symptoms (57.5%), pulmonal symptoms (37.2%) and pain symptoms (15.9%) 182 (Supplementary Table 1). Significantly more patients (18%) than donors (3%) needed oxygen (p=0.0014). 183 Hospitalization for any cause occurred in 7% of donors and in 18% of patients during the extended follow-184 up period (p=0.051). 185 The proportion of donors with functional limitations assessed by the post-COVID-Scale was lower than 186 the proportion in patients (22% vs. 56%, p<0.001), and correspondingly, the subgroup with grade 2-4 187 limitations was also lower in donors (10.6% vs. 48%, p<0.001) (Fig. 4 A). 188 189

Quality of life 190
A substantial proportion of patients (24%) reported a decrease in their socioeconomic status during 191 follow-up with only slight numerical difference between CCP group and control group patients (26.7% vs. 192 20.0%, p=0.74) (Fig. 4B). 193 Figure 5 shows a summary of total scores of the reported quality of life questionnaires. In the EQ 5D 5L 194 questionnaire the patients of the CCP group reported numerically better outcomes than the control group 195 in all five dimensions, i.e. mobility, self care, usual activities, pain/discomfort and anxiety (Supplemental 196 Table 2). The dimensions self-care, usual activities, pain/discomfort, anxiety and your health today were 197 not statistically different between CCP and control group, while a significantly higher proportion of 198 patients of the CCP than the control group reported that they have no problems in walking about (63% 199 vs. 40%, p=0.0395)(Supplemental Table 3). There was no relevant difference of the EQ-5D-5L items 200 between the patients of the low and high titer CCP group ( Figure 5A; Supplemental Table 4). The results 201 of the FACIT-dyspnea and FACIT-fatigue questionnaires show similar patterns: Scores were numerically 202 better in the CCP group than the control group without reaching statistical significance (Fig 5 B; 203 Supplemental Table 6 and 9). The difference between subgroups by cumulative amount of neutralizing 204 antibodies was small with a consistent trend for better scores in most of the items in the subgroup that 205 has received a higher cumulative amount of neutralizing units ( Figure 5B, Supplemental Table 7 and 10). 206 FACIT-Fatigue score and the individual items did not differ significantly in the comparisons by 207 randomization group (Supplemental Table 12) and by cumulative amount of transfused units 208 (Supplemental Table 14). 209 Significantly more patients (24%) than donors (2.7%) reported a decrease in their socioeconomic status 210 during follow-up (p<0.0001) (Fig.4B). In the EQ-5D-5L questionnaire donors reported significant better 211 outcomes in all five dimensions than patients (Supplemental Table 2). The visual scale score of the item 212 "your health today" was significantly higher in donors than in patients (p<0.0001)( Figure 5A; 213 Supplemental Table 2). 214 In all the quality of life questionnaires used in this study the donors showed significant better results 215 The score of the FACIT-Fatigue scale was significantly higher in donors than patients, indicating less fatigue 217 in the donor group (p=0.0038) (Fig. 5C, Supplemental Table 11). The majority of items, in particular "I have  218   energy", "I am able to do my usual activities", "I am too tired to eat", "I need help doing my usual activities"  219   and "I have to limit my social activity because I am tired" indicate significantly greater impairment in the  220   patient population (Supplemental Table 11). 221 Because of the differences between the donor and patient population the outcomes might be influenced 222 by other factors than severity of COVID-19. We therefore identified 26 pairs of donors and patients by 223 propensity score matching for the variables age, sex and BMI (Supplemental Table 15). In this matched 224 cohort the differences between donors and patients were significant for the change of socioeconomic 225 status and the Post-COVID-Scale (Supplemental Figure 3A and 3B), the EQ-ED-5DL visual scale and EQ-ED-226 5DL cross walk score (Supplemental Fig.4A), and the FACIT Dyspnea 2 score (Supplemental Figure 4B). 227 FACIT Fatigue and FACIT Dyspnea 1 score did not significantly differ between patients and donors in the 228 propensity score matched groups (Supplemental Fig.5B Figure 6 shows the results of the 237 neutralizing titer causing 50% inhibition in plaque-reduction neutralization test (PRNT50) at baseline or 238 first apheresis and after the long-term follow-up. Among vaccinated participants with available baseline 239 and follow-up data, patient PRNT50 titers increased from 1:80 (1:20-1:480) to 1:5120 (1:3840-1:5120). A 240 significant increase of PRNT50 titers from 1:80 (1:20-1:320) to 1:5120 (1:1600-1:5120) was also observed 241 in the patients randomized to CCP (p<0.0001)( Figure 6C). 242 Anti-SARS-CoV-2 IgG and IgA antibodies measured by ELISA increased significantly after vaccination of 243 patients. The use of CCP seems to have no effect on the increase of IgG or IgA by vaccination 244 (Supplemental Figure 2C). 245 Baseline PRNT50-titers in patients (1:120 (1:40 -1:320)) were significantly higher than in donors (1:80 246 (1:20-1:160)) (p=0.045) ( Figure 6A). Donor PRNT50 titers increased from 1:80 (1:20-1:160) to 1:2500 247 (1:1280 to 1:5120)( Figure 6A). Vaccinated patients had significantly higher PRNT50 values at follow-up 248 than vaccinated donors (p=0.0005) ( Figure 6B). 249 The baseline anti-SARS-CoV-2 IgG ratio (measured by ELISA) of donors (3.8 (2.9-5.8)) was comparable to 250 that of patients (3.4 (2.2-6.6)) (p=0.5), while the baseline IgA ratio was significantly higher in patients (7.0 251 (2.2-9.0)) compared to donors (2.3 (1.3-3.9)) (p<0.0001) (Supplemental Figure 2A). At last follow-up 252 vaccinated patients and donors had significantly higher IgG and IgA ratios compared to their respective 253 baseline ratios and IgG and also IgA ratios did not significantly differ between donors and patients at last 254 follow-up (Supplemental Figure 2B).  Table 14). 261

262
To the best of our knowledge, this is the first randomized clinical trial that reports long-term data on the 263 use of CCP with a median follow-up of more than 1 year. While many trials of CCP for COVID-19 at different 264 stages of COVID-19 have been published, they report on short observation periods, often just up to about 265 1 months or less after randomization (2-19). It is evident that during the pandemic, it was important to 266 make the initial results of the trials publicly available as soon as possible. However, the long-term results 267 must also be taken into account -especially as it became clear that long-term complications involving 268 different organ systems after COVID-19 are very common, significantly affect patients' quality of life and 269 also influence overall survival (26-30). 270 The risk of Long COVID-19 increases with age, pre-existing conditions and severity of COVID-19 (32-35). 271 Patients who had to be treated in hospital or patients who required intensive care have a higher risk of 272 Long COVID-19 than patients with a mild to moderate course who could be treated on an outpatient basis 273 (32,33). Thus, the risk for the manifestation of a long COVID-19 is also increased in the patients in the 274 CAPSID study: The median age in the study was 60 years, all cases had severe COVID-19 and a high 275 proportion of patients (89%) had a previous disease associated with an unfavorable course of  Thus, the study population of the CAPSID study represents a group of patients who are particularly at risk 277 for Long COVID-19 and who require follow-up for medical reasons.
The lack of knowledge applies also to CCP donors: Less is known about the long-term course of former 279 CCP donors. Therefore we included CCP donors in this analysis to learn more about their long-term disease 280 burden. They also comprised an additional reference group since they had experienced an asymptomatic 281 to moderate COVID-19 disease as opposed to the CAPSID trial patients who had severe COVIC-19. Results 282 of the CAPSID trial based on the initial 2-month observation period and the CCP donor characteristics have 283 been previously published (1,31). There are several factors which might influence long-term outcome: At 284 the time of the previous analysis not all patients had reached the respective endpoints (clinical 285 improvement, time to discharge from ICU and hospital). Given the burden of long-COVID-19 and persisting 286 organ dysfunction the outcome might change due to long-term sequela. The enrollment to the CAPSID 287 trial was completed a few days prior to availability of SARS-CoV-2 vaccines in Germany. Also new variants 288 evolved thereafter. It was not clear how vaccination and potential re-infections will impact the long-term 289 course. Therefore, we considered an extended follow-up necessary. Here we now provide an update 290 based on a median follow-up of 392 days. 291 The follow-up demonstrated a long-term OS which was numerically higher in the CCP group compared to 292 the control group -but the difference was not statistically significant. A pre-defined subgroups analysis 293 of the initial 2-months observation periods showed a significant benefit of CCP among patients who 294 received a higher amount of neutralizing antibodies (1). The significant effect of transfusion of a larger 295 amount of neutralizing units tended to be even more pronounced in the long-term observation across 296 several endpoints. In the previous report the day-60 probability of survival was 91.6% in the subgroup 297 that received a higher cumulative amount of neutralizing antibodies and 68.1% in the control group 298 CoV-2 RBD mutations, including those found in variants of concern (51). Repeated antigen exposure can 330 confer potency, breadth, and resilience to viral escape mutations (56). Therefore, for future CCP 331 programs, priority should be given to superimmunized donors with very high antibody concentration due 332 to previous SARS-CoV-2 infection and vaccination (54,55,57). 333 We used several instruments to assess QoL of donors and patients during the The CCP group and especially the subgroup that received a higher cumulative amount of neutralizing 345 antibodies showed consistently numerically better results but the differences did not reach statistical 346 significance for the individual item with the exception of the lower hospitalization rate in the high dose 347 subgroup. Nevertheless, the trend for better for less constraints in the CCP group was very consistent 348 across three different QoL instruments which cover different dimensions (Supplemental Tables 3, 6, 7 and  349   8). Also, the proportion of patients without pulmonary symptoms was lower in the CCP group compared 350 to the control group (53% vs. 30%), together with a lower proportion of patients with need for any type 351 of ventilation support during follow-up after the initial observation period in the CCP group compared to 352 the control group (20% vs. 60%). This might suggest less pulmonary impairment in the CCP group during 353 the extended follow-up period. 354 The frequency of Long COVID-19 varies greatly in the literature and ranges up to a proportion of over 80% 355 of patients who report at least one Long COVID-19 symptom (26,58,59). Common symptoms of Long 356 COVID-19 are fatigue (98%), myalgias (87%), headache (83%) and dyspnea (88%) (COVERSCAN study data, 357 based on patients with persistent symptoms) (58). Organs whose function may be impaired in Long 358 COVID-19 include lungs, heart, liver, kidneys and nervous system (29,33,58,60). The COVERSCAN study 359 reported that 70% of patients with Long COVID-19 symptoms still had impairment in at least one organ 360 system at least 4 months after acute COVID-19 disease (58). In a large cohort study from Wuhan, patients 361 reported mainly fatigue and muscle weakness (63%), sleep disturbance (26%), and anxiety and depression 362 (23%) after a median time of 176 days (34). Pulmonary diffusion disorders were detectable during follow-363 up of 56% of patients with WHO grade 5 or 6 COVID-19 (34) we cannot rule out that the difference is due to other confounding variables that might influence antibody 380 levels, e.g. age, BMI, or a different timing of immunization events. Patients are significantly older and their 381 interval since last vaccination and antibody measurement was longer than in donors (Table 1). There has been the concern that CCP treatment might impair response to vaccination later on (69). Our limited data 383 set does not support this notion. This aspect needs further investigation as we continue to use and design 384 antibody-based therapies for COVID-19 and other infectious diseases. 385 The main shortcoming of our study is the limited sample size that included only 50 patients into the long-386 term follow-up. The CAPSID trial treated patients with severe COVID-19. More than 50% of patients 387 included in this long-term observation period had a baseline WHO score of 5 or higher and the duration 388 from symptom start to randomization was 8 (5.0 -11.0) days. Meanwhile there are trials and registry 389 studies that suggest a higher efficacy of CCP when it is given early in the course of COVID-19 to patients 390 with mild symptoms (10,18,36,70,71). Therefore, the long-term effect of CCP might be too subtle in this 391 small cohort which represents a subgroup of patients with poor prognosis due to advanced disease and 392 late CCP treatment. The small sample size limits also a more detailed analysis of quality of life and 393 antibody-responses in the subgroups treated with low or high amounts of neutralizing antibodies. 394 Nevertheless, these data can provide a reference for the long-term burden of disease in patients treated 395 in a CCP trial -in particular since several validated and internationally widely used QoL instruments have 396 been used and a reference cohort of patients with mild to moderate disease (donors) was included. 397 In conclusion, the consistent trend for a benefit across several endpoints (overall survival, time to first 398 negative SARS-CoV-2 PCR, discharge from ICU, discharge from hospital) among patients who received a 399 larger cumulative amount of neutralizing antibodies is confirmed in the extended observation period. One outpatient visit between day 240 and 540 after randomization or first plasma donation was planned. 430 The following assessments and data collections were performed: medical history including symptoms, 431 complications, hospital treatments, medication and chest imaging since the previous end of study, 432 heartrate and blood pressure, Quality of life-questionnaires (EDQ5, FACIT fatigue and FACIT Dyspnea), 433 and blood tests for inflammation markers, coagulation markers, anti-SARS-CoV-2 immunity and organ 434 function. A structured interview was performed using a pre-specified questionnaire and the long COVID-435

Scale (72). 436
Patients who could not visit the study center could also participate by telephone. In these cases, no 437 laboratory values were collected and no functional tests were performed. 438

SARS-CoV-2-Antibody Assays 440
A plaque reduction neutralization test (PRNT) and an ELISA for the detection of IgG and IgA to the S protein 441 of SARS-CoV-2 were performed as previously described (1,(73)(74)(75). The 5Q-5D-5L questionnaire assesses 5 dimensions: mobility, self care, usual activities, pain/discomfort, 454 anxiety in 5 categories and one´s health today by a visual analogue scale giving an EQ-5D-5L index score 455 (76). FACIT fatigue and FACIT Dyspnea were also used. The FACIT fatigue questionnaire, which consists of 456 13 questions, was originally developed to understand the impact of anemia and fatigue on the daily 457 activities of cancer patients, but it has also been used for many other chronic diseases (77). For each 458 question, there are 5 response options depending on the severity from "not at all" to "frequently." The 459 total score is on a numerical scale from 0 to 52, whereby the higher the score, the lesser the fatigue.